Rational Vs Adaptive Exectations

I really shouldn’t comment on Simon Wren_lewis’s defence of rational expectations until I have calmed down, but I can’t help muself. I will try to stick to FRED, that is data.

Wren-Lewis argues that it is reasonable for macroeconomists to assume rational expectations since the practical alternatives are rational expectations or naive expectations.

However most of the time macroeconomists want to focus on something else, and so we need a simpler framework. In practice that seems to me to involve a binary choice. Either we assume that agents are very naive, and adopt something very simple like adaptive expectations (inflation tomorrow will be based on current and past inflation), or we assume rational expectations. My suspicion is that heterodox economists, when they do practical macroeconomics, adopt the assumption that expectations are naive, if they exist at all (e.g. here). So I want to explain why, most of the time, this is the wrong choice. My argument here is similar but complementary to a recent piece by Mark Thoma on rational expectations.

Suppose we have an equation determining wage or price inflation (a Phillips curve), where inflation expectations appear on the right hand side of the equation. We need some way of determining those expectations. Lots of nice words like ‘non-ergodic’ will not do: we need something simple that can be used to solve the model. To assume, as mainstream macroeconomists once did, that these expectations just depend on past observations about inflation seems to assume that agents are stupid. These agents ignore everything that economists and the media say about inflation: they ignore monetary policy, and whether the economy is in a boom or recession.

I will assume for the sake of argument that this is true. I think it is obvous that it is better to assume adaptive than rational expectaions when attempting to model advanced industrial economies and to guide policy.

I note that the assumption of naive expectations leads to the belief that there will be irrational speculative bubbles in which agents assume some asset price will increase because it has in the past. This is one of they key features of the data. It is possible to reconcile this witih the rational expectations assumption, because anything at all can be reconciled with the assumption (note I never assert that the rational expectations hypothesis is false since we all agree that there is no falsifiable rational expectations hypothesis).

The rational expectations assumption convinced policy makers in the early 80s that disinflation would not be costly provided that iron resolve etc gave the dry policy credibility. It was exceedingly costly.

But mostly I think that the rational vs adaptive expectations debate can best be addressed just by looking at a bit of data which is not decisive but, to me, convincing on the order of the anthropogenic global warming debate.

I ask Provessor Wren-Lewis how often he looks at a graph like the one below (made at FRED). I am askkng for information and I hope to get an answer.


The blue curve is inflation expected by ordinary people from the Michigan Survey, the red curve is cpi inflation in the yer till the date of the survey and the green line is core inflation (CPI minus food and energy). It seems to me that survey expected inflation is equal to CPI inflation except when CPI infation is extraordinarily different from core inflation because of a recent Iraqi invasion of Iran or theh 2008 collapse of demand for, among other things, petroleum. This simple ad hoc model CPI or core if very different fits the data rather well. I would say its empricial success is vastly greater than the empirical success of any micro founded macro model..I note that my model is much more naive than the paleo-Keynesian approach. Adaptive expectations are a weighted sum of the old expected inflation and lagged actual inflation. They smooth off peaks as in 1980 and early 2009. The blue curve looks like the red curve smoothed is a pretty good summary of data all of which was collected *after* rational expectations assumption was declared the winner of the debate.

Thre is another anomaly. In recent years Michigan survey forecast inflation is persistently higher than actual inflation. At the same time the general public’s estimates of achieved inflation are higher than official calculations. The rational expectations assumption has very strong implications for statements about data available at the time the statement was made. These implications are totally rejected by the data.

OK back to Wren-Lewis’s critique of the adaptive expectations hypothesis

“These agents ignore everything that economists and the media say about inflation: they ignore monetary policy, and whether the economy is in a boom or recession” Most economic agents in the USA clearly ignore what economists and the media say about inflation in the past few years. This is plainly true. Wren Lewis knows it is true (he wrote so in another post). So why is it better to assume that people read, here *an believe* something when they insist that they don’t ? Not acceptable as a useful approximation but necessary no matter how well another approaach fits the data. How about shifts in monetary policy. Do you think anyone who didn’t know when monetary policy shifted could figure out the dates by contrasting the blue and red lines ? Is there any hint of a trace of evidence in the data that a huge tightening of monetary policy causes expected inflation to be lower than one would guess using only lagged inflation ? I ure don’t see it. And the behavior in recessions is completely different in different recessionss.

All of WRen-Lewis’s question are easily answered with the most cursory glance at the relevant data. The answer, as always for developed countries, is that the evidemce supports the assumption of adaptive expectations.

Now this doesn’t mean that it is reasonable to assume adaptive expectations when considering hyperinflation. Long ag a macroeconomist correctly noted that there aren’t simple patterns relating real and nominal variables and, in particular, everything is close to neoclassical in the case of hyperinflation. Long ago means 1936 and the economist is Keynes. I am. as usual araphrasing The General Theory… If a macroeconomist suggested in 1968 that this was a new insight, then he was lying.

OK post too long. I have tried to avoid being rude (really) and see an earlier post for more recent data supporting the adaptive expectations hypothesis.

Comments (9) | |